Modeling spatial covariation of summer temperatures and bio-indicators in an Arctic coastal area

• Published in: Climate Research
• Main Authors: Nilsen, Lennart, Joly, Daniel, Elvebakk, Arve, Brossard, Thierry
• Other Authors: University of Tromsø (UiT), Théoriser et modéliser pour aménager (UMR 6049) (ThéMA), Université de Bourgogne (UB)-Centre National de la Recherche Scientifique (CNRS)-Université de Franche-Comté (UFC), Université Bourgogne Franche-Comté COMUE (UBFC)-Université Bourgogne Franche-Comté COMUE (UBFC)
• Format: Article in Journal/Newspaper
• Language: English
• Published: HAL CCSD 2013
• Subjects: Svalbard; bio-indicators; rRegression analysis; Temperature; vegetation; [SHS.GEO]Humanities and Social Sciences/Geography; Arctic
• Online Access: https://hal.archives-ouvertes.fr/hal-00911237; https://doi.org/10.3354/cr01173

International audience In the Arctic, temperature is a major environmental factor controlling the occurrence, abundance and distribution of plants at regional and local scales alike. This means that statistical models of temperature distribution can predict the distribution of plant species or communities. Conversely, certain plant taxa make good bio-indicators reflecting long-term thermal conditions in a given habitat. Both these assumptions were taken into account when modelling the spatial relationship between plants and temperature. This work continues a previous preliminary 1 yr study based on correlations between a plant-based index of thermophily (It) and different synthetic temperature distribution characteristics. To strengthen confidence in the results and conclusions, more temperature data were collected through a field campaign conducted over a further 5 yr period (2001 to 2005). The goals here were (1) to establish an accurate interpolation model capable of restoring, at local scale, the continuous summertime thermal raster surface, (2) to evaluate the capacity of the temperature values obtained from the model to predict the distribution of It, and (3) to extrapolate temperature surfaces from this It. The results show that the mutual predictive power between temperature and It is satisfactory and that the model can be applied to neighbouring areas, although the present study area is too small to define the geographical limits of extrapolation. This predictive power declines where local landscape structures are heterogeneous. Correlations between It and growing degree day (GDD) values derived from the modelled temperature layers were systematically analysed in order to identify conditions in which this covariation works or fails.